Content uploaded by Roger F L James
Author content
All content in this area was uploaded by Roger F L James on Nov 15, 2016
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
Glucose Stimulates Translocation of the Homeodomain
Transcription Factor PDX1 from the Cytoplasm to the
Nucleus in Pancreatic
b
-Cells*
(Received for publication, June 11, 1998, and in revised form, October 7, 1998)
Wendy M. Macfarlane‡, Caroline M. McKinnon‡, Zoe A. Felton-Edkins‡, Helen Cragg‡,
Roger F. L. James§, and Kevin Docherty‡¶
From the ‡Department of Molecular and Cell Biology, Institute of Medical Sciences, Foresterhill, Aberdeen AB25 2ZD and
the §Department of Surgery, University of Leicester, Leicester Royal Infirmary, Leicester LE2 7LX, United Kingdom
One of the mechanisms whereby glucose stimulates
insulin gene transcription in pancreatic
b
-cells involves
activation of the homeodomain transcription factor
PDX1 (pancreatic/duodenal homeobox-1) via a stress-
activated pathway involving stress-activated protein ki-
nase 2 (SAPK2, also termed RK/p38, CSBP, and Mxi2). In
the present study we show, by Western blotting and
electrophoretic mobility shift assay, that in human islets
of Langerhans incubated in low glucose (3 mM) PDX1
exists as an inactive 31-kDa protein localized exclu-
sively in the cytoplasm. Transfer of the islets to high (16
mM) glucose results in rapid (within 10 min) conversion
of PDX1 to an active 46-kDa form that was present pre-
dominantly in the nucleus. Activation of PDX1 appears
to involve phosphorylation, as shown by incorporation
of
32
P
i
into the 46-kDa form of the protein. These effects
of glucose could be mimicked by chemical stress (sodi-
um arsenite), or by overexpression of SAPK2 in the
b
-cell line MIN6. Overexpression of SAPK2 also stimu-
lated PDX1-dependent transcription of a –50 to –250 re-
gion of the human insulin gene promoter linked to a
firefly luciferase reporter gene. The effects of glucose
were inhibited by the SAPK2 inhibitor SB 203580, and by
wortmannin and LY 294002, which inhibit phosphatidyl-
inositol 3-kinase, although the effects of stress (arsenite)
were inhibited only by SB 203580. These results demon-
strate that glucose regulates the insulin gene promoter
through activation and nuclear translocation of PDX1
via the SAPK2 pathway.
The homeodomain transcription factor PDX1 (pancreatic/
duodenal homeobox-1), which has previously been called IPF1
(1), IDX1 (2), STF1 (3), or IUF1 (4), plays an important role in
lineage determination in the developing endocrine pancreas (5,
6). PDX1 binds to four sites (A1, A2, A3, and A5) with the
consensus sequence C(C/T)TAATG in the human insulin pro-
moter (7). Although it can transactivate the insulin promoter
(8–10), its exact role in basal insulin transcription is unclear
since insulin gene transcription can occur in the absence of
PDX1 (11, 12). However, there is strong evidence supporting a
role for PDX1 in the mechanisms whereby glucose stimulates
insulin gene transcription in response to changes in glucose
concentrations (13–15). Furthermore, glucose metabolism has
been shown to stimulate PDX1 DNA binding activity and in-
sulin promoter activity in
b
-cells via a stress-activated signal-
ing pathway involving stress-activated protein kinase 2
(SAPK2, also termed RK/p38, CSBP, and Mxi2) (16).
SAPK2 is a member of an expanding family of mitogen-
activated protein kinase-related kinases that are activated in
response to adverse stimuli such as heat, osmotic shock, ultra-
violet light, DNA-damaging reagents and by proinflammatory
cytokines that are produced under conditions of stress (17). In
pancreatic
b
-cells, glucose metabolism stimulates the SAPK2
pathway by a mechanism that involves phosphatidylinositol
3-kinase (PI 3-kinase),
1
while stress (e.g. arsenite treatment)
stimulates the SAPK2 pathway in
b
and other cell types by a
mechanism independent of PI 3-kinase (16).
Recombinant PDX1, synthesized in Escherichia coli, has a
molecular mass of 31 kDa, similar to that predicted from its
amino acid sequence. It is inactive, insofar as it does not bind to
its recognition sequence in the insulin promoter as measured
by electrophoretic mobility shift assay. However, treatment of
recombinant PDX1 with SAPK2 in the presence of Mg-ATP and
a
b
-cell extract activates its DNA binding, concomitant with a
shift in molecular mass from 31 kDa to 46 kDa (16). The
molecular modifications responsible for this shift in molecular
mass are as yet unclear. In the present study we show that
PDX1 exists as the inactive 31-kDa form in the cytoplasm of
human islets incubated in low (3 mM) glucose. Treatment with
16 mMglucose stimulates conversion to the active 46-kDa form,
which is localized predominantly in the nucleus. These events
are inhibited by SB 203580 (5
m
M) (an inhibitor of SAPK2),
wortmannin, and LY 294002 (inhibitors of PI 3-kinase), and
mimicked by sodium arsenite and overexpression of SAPK2.
EXPERIMENTAL PROCEDURES
Chemicals and Reagents—Radiochemicals were purchased from Am-
ersham International (Slough, Berks, United Kingdom (UK)), sodium
arsenite from Fisons (Loughborough, UK), and wortmannin from Sigma
(Poole, UK). SB 203580 was a generous gift from Dr. J. Lee and Dr. P.
Young (SmithKline Beecham, King of Prussia, PA). Anti-PDX1 anti-
body was a kind gift from Dr. C. V. Wright (Vanderbilt University,
Nashville, TN). Anti-SAPK2 antibody was purchased from New Eng-
land Biolabs (Cambridge, UK). Anti-upstream stimulatory factor (USF)
antibody was kindly provided by Dr. M. Sawadogo (University of Texas,
Houston, TX).
Oligonucleotides—Oligodeoxynucleotides were purchased from Alta
Bioscience (University of Birmingham, Birmingham, UK). Single-
stranded complementary oligonucleotides were annealed as described
* This work was supported by grants from the Wellcome Trust and
the British Diabetic Association and by a studentship (to C. M. M.) from
the British Diabetic Association. The costs of publication of this article
were defrayed in part by the payment of page charges. This article must
therefore be hereby marked “advertisement” in accordance with 18
U.S.C. Section 1734 solely to indicate this fact.
¶To whom correspondence should be addressed. Fax: 44-1224-
273144.
1
The abbreviations used are: PI 3-kinase, phosphatidylinositol 3-ki-
nase; USF, upstream stimulatory factor; EMSA, electrophoretic mobil-
ity shift assay; PAGE, polyacrylamide gel electrophoresis; IL, interleu-
kin; NF, nuclear factor.
THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 274, No. 2, Issue of January 8, pp. 1011–1016, 1999
© 1999 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.
This paper is available on line at http://www.jbc.org 1011
by guest on November 15, 2016http://www.jbc.org/Downloaded from
previously (18), and labeled with [
g
-
32
P]ATP using T4 polynucleotide
kinase. Oligonucleotide B, corresponding to the A3 site of the human
insulin gene promoter, is a complementary double-stranded 30-mer
59-CCCCTGGTTAAGACTCTAATGACCCGCTGG-39.
Isolation and Treatment of Human Islets of Langerhans—Human
islets of Langerhans were isolated from the pancreas of human organ
donors, and all procedures were carried out with the approval of the
appropriate Ethical Committee. The pancreatic duct was cannulated in
situ and digestion achieved by intraductal infusion of collagenase. The
islets were then separated from contaminating acinar tissue by centrif-
ugation on discontinuous density gradients of bovine serum albumin in
a semi-automated system (19). The purified islets were placed in RPMI
1640 medium (Life Technologies, Inc.) containing 10% (v/v) fetal calf
serum and supplemented with 400 IU/ml sodium penicillin G and 200
m
g/ml streptomycin sulfate, and cultured at 37 °C in a humidified
atmosphere of O
2
:CO
2
(95:5) for several days prior to use. Selected islets
were separated into batches of 120–150 in Hanks’ buffered saline (0.12
MNaCl, 5.4 mMKCl, 0.33 mMNa
2
HPO
4
, 0.44 mMKH
2
PO
4
, 0.4 mM
MgSO
4
, 0.5 mMMgCl
2
,1mMCaCl
2
,2mMNaHCO
3
, 5% (w/v) bovine
serum albumin) except where specified.
Preparation of Cell Extracts—Nuclear and cytoplasmic extracts were
prepared using a modification of the method of Schreiber et al. (20).
Cells were centrifuged for 10 s in a microcentrifuge, and resuspended in
400
m
lof10mMHepes, pH 7.9, containing 10 mMKCl, 0.1 mMEDTA,
0.1 mMEGTA, 1 mMdithiothreitol, 0.5 mMphenylmethanesulfonyl
fluoride, 10 mMNaF, 10 mMsodium molybdate, 10 mM
b
-glycerophos-
phate, 10 mMsodium vanadate, and 10 mMp-nitrophenyl phosphate.
Cells were allowed to swell on ice for 15 min before adding 25
m
lof10%
(v/v) Nonidet P-40. The islets were then incubated for an additional 30
min on ice, vortexed for 15 s, and centrifuged for 30 s in a microcentri-
fuge. The cytoplasmic fraction (supernatant) was removed and frozen in
small aliquots, and the pellet, which was enriched in nuclei, was resus-
pended in 50
m
lof20mMHepes, pH 7.9, containing 0.4 MNaCl, 1 mM
EDTA, 1 mMEGTA, 1 mMdithiothreitol, 1 mMphenylmethanesulfonyl
fluoride, 10
m
g/ml leupeptin, 1
m
g/ml pepstatin A, 0.1 mMp-aminoben-
zoic acid, 10
m
g/ml aprotinin, 5% (v/v) glycerol, 10 mMNaF, 10 mM
sodium molybdate, 10 mM
b
-glycerophosphate, 10 mMsodium vanadate,
and 10 mMp-nitrophenyl phosphate. Nuclear extracts were then cen-
trifuged for 2 min at 4 °C in a microcentrifuge. The supernatant was
collected, aliquoted into small volumes, and stored at 270 °C. For whole
cell extracts, Nonidet P-40 lysis was replaced with incubation with 1%
(v/v) Triton X-100, vortexing for 1 min, and centrifugation for 2 min.
Electrophoretic Mobility Shift Assays (EMSAs)—EMSAs were per-
formed as described previously (16). Cell extracts (0.5
m
g of protein)
were incubated with radiolabeled probe for 20 min at room temperature
in buffer containing 10 mMTris-HCl, pH 7.5, 50 mMKCl, 5 mMdithio-
threitol, 1 mMEDTA, and 5% (v/v) glycerol.
Cell Culture—MIN6 cells were grown in Dulbecco’s modified Eagle’s
medium containing 5 mMglucose, supplemented with 15% heat-inacti-
vated myoclone fetal calf serum (Sigma) and 2 mML-glutamine, in a
humidified atmosphere containing 95% air, 5% CO
2
.
Plasmids—The control construct pGL-LUC is based on the plasmid
pGL2 (Promega), with the thymidine kinase promoter from the herpes
simplex virus cloned 59to the firefly luciferase gene. The construct
pGL-LUC200 varies from this in that it contains 200 base pairs of the
human insulin gene promotor. A 260 to 2260 base pair HincII-PvuII
fragment from the human insulin gene promoter was blunt-ended and
cloned into the SmaI site of the control construct. DNA was prepared
using the Qiagen Endotoxin-Free Maxiprep method, and quantitated
spectrophotometrically. PDX1 expression vector pCR3-PDX1 was as
described previously (11). The pSG5-SAPK2 (21) expression vector was
a generous gift from Dr. S. Keyse, University of Dundee.
Transfections—Cells were grown to about 80% confluence in six-well
plates, and were transfected by mixing 4
m
g of DNA and 54
m
lofa1nM
lipid suspension containing a 2:1 mixture of dioleoyl-L-
a
-phosphati-
dylethanolamine (Sigma) and dimethyl-dioctadecylammonium bromide
(Fluka) in 1 ml of serum free Opti-MEM (Life Technologies, Inc.). The
lipid-DNA complexes were allowed to form for 20 min at room temper-
ature before being added to the washed cells. Following5hofincuba-
tion, 1 ml of complete medium containing 30% heat-inactivated
myoclone fetal calf serum was added to the cells. After 12 h, the
medium-DNA complexes were replaced by complete medium and the
cells left for another 24 h before treating. Treatment of all cells started
with a 5-h preincubation in Dulbecco’s modified Eagle’s medium con-
taining 0.5 mMglucose. Cells were then removed from the wells and a
cell pellet recovered by centrifugation at 7000 rpm for 30 s. The cell
pellet was resuspended in 70
m
l of 100 mMKH
2
PO
4
,pH7.8,1mM
dithiothreitol solution and lysed by freeze/thawing three times. Cell
debris was removed by centrifugation at 13,000 rpm for 1 min.
Luciferase Assay—30
m
l of cell extract was added to 350
m
l of buffer
A, pH 7.8 (15 mMMgSO
4
,30mMglycylglycine, 2 mMNa
2
ATP) contain-
ing 0.45 mMcoenzyme A and 2.56 mMTriton X-100. To this, 150
m
lof
buffer G (30 mMglycylglycine) containing 0.5 mMluciferin (Sigma) was
injected and the luminescence read at 560 nMusing a Berthold Luma
LB9501. Protein content of the cell extract was measured by taking 6
m
l
of extract and adding144
m
l of water and 150
m
l of Coomassie protein
assay reagent (Pierce). Protein content was read at a wavelength of 620
nm using a Titer Tech Multiscan MCC340. A standard curve of known
bovine serum albumin protein concentrations was used to calculate the
protein concentrations of the samples.
Western Blotting—For Western blot analysis, 1-
m
g samples of cell
extract were fractionated by SDS-PAGE, blotted on to ECL-nitrocellu-
lose membrane (Amersham), and incubated for 60 min in a buffer
containing 10 mMTris-HCl, 0.05% (v/v) Tween 20, 0.5 MNaCl, and a
1:5000 dilution of anti-PDX1 antibody. The antigen-antibody complex
was then detected by incubating the membrane for another 60 min in
buffer containing a 1:5000 dilution of horseradish peroxidase-conju-
gated anti-rabbit IgG secondary antibody (ECL, Amersham).
32
P
i
Labeling and PDX1 Immunoprecipitation—MIN6 cells were
washed three times in Krebs Ringer (118 mMNaCl, 4.75 mMKCl, 1.2
mMMgCl
2
, 0.26 mMCaCl
2
,25mMNaHCO
3
) prior to addition of 0.5 mCi
of
32
P
i
/10
6
cells in Krebs Ringer. Cells were incubated at 37 °C for 1 h,
prior to addition of glucose to 0.5 or 16 mMfinal concentration. After 30
min cells were washed three times in Krebs Ringer and harvested, with
nuclear and cytoplasmic extracts prepared as before. For PDX1 immu-
noprecipitation: 400-
m
l samples in 13NDET (1% Nonidet P-40, 0.4%
deoxycholate, 66 mMEDTA, 10 mMTris-HCl, pH 7.4) were precleared
by addition of 5
m
l of preimmune rabbit serum and 10
m
l of protein A (in
13NDET), with rotation at 4 °C for 2 h. Preclearing was completed by
centrifugation at 13,000 rpm for 2 min at 4 °C, with the resulting pellet
being discarded. 3
m
l of anti-PDX1 antibody and 10
m
l of protein A-
Sepharose were added to 400
m
l of supernatant, and PDX1 immuno-
precipitated by rotation at 4 °C for 16 h. A 2-min centrifugation yielded
a pellet, which was then resuspended in 0.5 ml of NDET 10.3% SDS,
with rotation at 4 °C for 2 h. Samples were then layered over 0.5 ml of
NDET 10.3% SDS/30% sucrose, and centrifuged for 5 min at 4 °C. The
resulting pellet was resuspended in 0.5 ml of NDET and rotated at 4 °C
for another 2 h. Finally, the samples were centrifuged for 4 min at 4 °C,
with the final pellet resuspended in 30
m
l of SDS-PAGE sample buffer
(4% SDS, 0.02% bromphenol blue, 0.5 Msucrose, 10 mMTris-HCl, pH
6.8, 10%
b
-mercaptoethanol). Samples were heated to 80 °C for 5 min
prior to loading.
RESULTS
Initial experiments were undertaken to determine the effect
of glucose on the molecular mass and intracellular location of
PDX1. Human islets of Langerhans were incubated for5hin3
mMglucose, transferred to 16 mMglucose, and PDX1 analyzed
at various time intervals by Western blotting of whole cell
extracts. In 3 mMglucose PDX1 was present as a 31-kDa
protein (Fig. 1A). Within 10 min in 16 mMglucose, the 31-kDa
protein was converted to the 46-kDa form. Conversion was
complete within 15 min. To investigate the effect of glucose on
the intracellular location of the two forms of PDX1, human
islets were incubated in 3 or 20 mMglucose and cytoplasmic
and nuclear fractions prepared. In 3 mMglucose, the 31-kDa
form was present in the cytoplasm; however, in 16 mMglucose,
the 46-kDa form was present in the cytoplasm and the nucleus
(Fig. 1B). Treatment with acid phosphatase converted the 46-
kDa form to the 31-kDa form. In the presence of excess inor-
ganic pyrophosphate, acid phosphatase had no effect on the
46-kDa form, confirming that conversion to the 31-kDa form
was associated with dephosphorylation of the 46-kDa form. The
31-kDa form was inactive as measured by EMSA (Fig. 1C,
lanes 1–4), whereas the 46-kDa form present in the cytoplasm
and the nucleus was active (Fig. 1C,lanes 4–8). Identification
of the retarded band in Fig. 1Cas PDX1 was confirmed by
oligonucleotide and antibody competition experiments (data
not shown).
The role of phosphorylation in the activation of PDX1 was
investigated further by labeling of MIN6 cells with inorganic
Nuclear Translocation of PDX11012
by guest on November 15, 2016http://www.jbc.org/Downloaded from
32
P
i
and subsequent immunoprecipitation of PDX1. At high
glucose concentrations,
32
P
i
was incorporated into the 46-kDa
form of PDX1 (Fig. 2, lanes 3 and 4), but at low glucose con-
centrations no such incorporation was observed (lanes 1 and 2).
Western blot analysis of the immunoprecipitated samples (Fig.
2B) confirmed that both forms of the protein were present, but
that
32
P was incorporated only into the 46-kDa form of the
protein, which was observed at high glucose concentrations. To
evaluate the effectiveness of the biochemical fractionation of
the
b
-cell extracts, nuclear and cytoplasmic fractions were an-
alyzed by Western blotting using a specific anti-USF antibody.
USF binds to the E2 site of the human insulin gene promoter
and has a molecular size of 43 kDa (Fig. 2C). Blotting con-
firmed the presence of USF solely in the nuclear fractions,
independently of glucose concentrations.
Since SAPK2 and PI 3-kinase had previously been impli-
cated in the stimulation of PDX1 DNA binding activity, the
possible involvement of these enzymes in the glucose-stimu-
lated translocation of PDX1 to the nucleus was next investi-
gated. Fig. 3Ashows that glucose-stimulated conversion of
PDX1 from the 31-kDa to the 46-kDa form and translocation of
the 46-kDa form into the nucleus was inhibited by the SAPK2
inhibitor SB 203580, and by the PI 3-kinase inhibitors wort-
mannin and LY 294002 (Fig. 3A). SAPK2 is a stress-activated
protein kinase, the activity of which can be stimulated by
treating cells with the stress-inducing agent sodium arsenite (1
mM). The effects of glucose on PDX1 molecular mass and intra-
cellular location were mimicked by sodium arsenite, further
supporting a role for SAPK2 in these events (Fig. 3B). As shown
previously (16) the effects of arsenite were inhibited by SB
203580 but not by wortmannin and LY 294002 (Fig. 3B). The
IC
50
for SB 203580 on both glucose and arsenite effects was 3
m
M(data not shown).
To investigate further the involvement of SAPK2, MIN6 cells
were transfected with a cDNA encoding SAPK2. Overexpres-
sion was confirmed by the appearance of a 38-kDa immunore-
active protein in extracts from transfected cells that was absent
in untransfected cells (Fig. 4A). In untransfected MIN6 cells
treated with 3 mMglucose, PDX1 was present as the 31-kDa
form in the cytoplasm. Following treatment with 16 mMglu-
cose, PDX1 was converted to the 46-kDa form, which was
present in the cytoplasm and nucleus. When cells transfected
with SAPK2 were incubated in 3 mMglucose, PDX1 was pres-
ent as the 46-kDa form in the cytoplasm and in the nucleus
(Fig. 4B). This result suggested that overexpression of SAPK2
resulted in conversion of PDX1 to the 46-kDa form that was
localized predominantly in the nucleus. Overexpression of
SAPK2 also activated PDX1 DNA binding activity in low glu-
cose-treated MIN6 cells (Fig. 4C). The critical events governing
modification and activation of PDX1 appear to occur in the
cytoplasm, as Western blot analysis of nuclear and cytoplasmic
extracts prepared from MIN6 cells overexpressing SAPK2
clearly shows that the SAPK2 protein is localized in the cyto-
plasm and is undetectable in the
b
-cell nucleus (Fig. 5A).
As shown previously (16), glucose stimulates the transcrip-
tional activity of a DNA construct containing the firefly lucif-
erase gene under the control of the thymidine kinase reporter,
which is driven by a human insulin gene fragment from 260 to
2260 (pGL -LUC200) (Fig. 5). Overexpression of SAPK2 in
MIN6 cells had no effect on the control vector pGL-LUC, which
lacked the insulin gene fragment, but gave a 5-fold stimulation
of pGL-LUC200 activity, similar to that seen with 16 mMglu-
cose (Fig. 5). The effect of overexpression of SAPK2 on the
LUC200 activity was inhibited by SB 203580 (10
m
M), confirm-
ing that these effects were the direct consequence of the higher
SAPK2 activity in the transfected cells.
DISCUSSION
We have previously shown that glucose activates PDX1 DNA
binding activity and insulin promoter activity in pancreatic
b
-cells via a pathway involving SAPK2 (RK/p38). We also
showed that SAPK2 in vitro activated recombinant PDX1 DNA
binding activity, concomitant with a change in the molecular
mass of PDX1 from 31 kDa to 46 kDa (16). In the present study,
we show that glucose stimulates the conversion of PDX1 from
a 31-kDa unphosphorylated to a 46-kDa phosphorylated form
in isolated human islets of Langerhans (Fig. 1) and in MIN6
cells (Fig. 4). The activation of PDX1, and translocation to the
nucleus, were inhibited by SB 203580 and stimulated by ar-
senite and overexpression of SAPK2. This strongly supports a
role for SAPK2 in these events. That glucose stimulates a rapid
translocation of PDX1 to the nucleus is compatible with the
results of a recent study showing that glucose stimulates trans-
location of a c-Myc-tagged PDX1 (IPF-1) from the nuclear pe-
riphery to the nucleoplasm (22).
SAPK2-dependent phosphorylation of PDX1 appears to rep-
FIG.1.PDX1 is rapidly modified by
high glucose concentrations in hu-
man islets of Langerhans. A, Western
blot analysis of human islets of Langer-
hans incubated in 3 mMglucose for 5 h,
then transferred to 20 mMglucose for the
time periods indicated. Whole cell ex-
tracts were prepared, and probed with a
specific anti-PDX1 antibody. B, Western
blot analysis of cytoplasmic (C) and nu-
clear (N) extracts prepared from human
islets of Langerhans incubated in 3 mM
glucose for 5 h, or 3 mMglucose for 5 h
followed by 20 mMglucose for 30 min. In
lane 5, high glucose nuclear extract was
incubated for 30 min with 10 units of po-
tato acid phosphatase (AP). In lane 6, acid
phosphatase treatment was carried out in
the presence of 50 mMpyrophosphate (P
i
).
C, electrophoretic mobility shift assay of
cytoplasmic (C) and nuclear (N) samples
prepared from human islets of Langer-
hans incubated in 3 mMglucose for 5 h, or
3m
Mglucose for 5 h followed by 20 mM
glucose for 30 min, using the A3 site of the
human insulin gene promoter as a probe.
Nuclear Translocation of PDX1 1013
by guest on November 15, 2016http://www.jbc.org/Downloaded from
resent the critical step in glucose-induced translocation to the
nucleus and activation of DNA binding activity. However, it is
unclear whether phosphorylation alone accounts for the ob-
served change in size. It is possible that phosphorylation allows
or promotes a second event, which is yet to be characterized.
Phosphorylation may induce a conformational change in PDX1
that affects its mobility in SDS-PAGE, but the change in size
could equally result from a further post-translational modifi-
cation. Many transcription factors contain nuclear localization
signals within their amino acid sequences, but others, such as
PDX1, do not contain recognizable nuclear localization signals
and may require modification at the post-translational level.
FIG.2. High glucose promotes in-
corporation of
32
P
i
into the 46-kDa
form of PDX1. A, SDS-PAGE analysis of
PDX1 immunoprecipitated from nuclear
(N) and cytoplasmic (C) extracts prepared
from MIN6 cells incubated for3hin0.5
mMglucose, then labeled with
32
P
i
. La-
beled cells were incubated in 0.5 mMglu-
cose (lanes 1 and 2)or16m
Mglucose
(lanes 3 and 4) for 30 min prior to the
preparation of extracts.
32
P-Labeled
PDX1 is observed in lanes 3 and 4at 46
kDa. B, Western blot analysis of immuno-
precipitates described in A, using a spe-
cific anti-PDX1 antibody. C, Western blot
analysis of the cytoplasmic (C) and nu-
clear (N) extracts from MIN6 cells incu-
bated in 0.5 mMglucose for 5 h, or incu-
bated in 0.5 mMglucose for 5 h followed by
stimulation in 16 mMglucose for 30 min,
using a specific anti-USF antibody. USF
has a molecular mass of 43 kDa.
FIG.3. High glucose and chemical
stress induce modification of PDX1.
A, Western blot (anti-PDX1) analysis of
cytoplasmic (C,lanes 3–6), and nuclear
(N,lanes 7–10) samples prepared from
human islets of Langerhans incubated in
3m
Mglucose for5h(lanes 1 and 2), or 3
mMglucose for 5 h followed by 20 mM
glucose for 30 min, or in 20 mMglucose for
30 min in the presence of the inhibitors
indicated. SB 203580 (SB,20
m
M)LY
294002 (LY,50
m
M), and wortmannin
(Wtm,50n
M) were added 30 min prior to
stimulation. B, Western blot (anti-PDX1)
analysis of cytoplasmic (C,lanes 3–6),
and nuclear (N,lanes 7–10) samples pre-
pared from human islets of Langerhans
incubated in 3 mMglucose for5h(lanes 1
and 2), or in 3 mMglucose for 5 h followed
by1m
Msodium arsenite for 30 min in the
presence of the inhibitors indicated.
SB203580 (SB,20
m
M) LY294002 (LY,50
m
M), and wortmannin (Wtm,50nM) were
added 30 min prior to stimulation.
Nuclear Translocation of PDX11014
by guest on November 15, 2016http://www.jbc.org/Downloaded from
Common modifications promoting nuclear localization include
glycosylation (23), and ubiquitination (24, 25), which modifies
proteins through the addition of ubiquitin-like protein species
such as SUMO1 (small ubiquitin-related modifier 1). In the
case of RanGAP (Ran-GTPase activating protein) (25), phos-
phorylation allows the addition of SUMO1, which alters the
apparent molecular mass of the protein by 20 kDa, and pro-
motes its translocation from the cytoplasm to the nucleus.
Similar events may be occurring with PDX1.
Increased binding activity of transcription factors through
phosphorylation can occur by several mechanisms. Many tran-
scription factors become phosphorylated in the cytoplasm,
stimulating movement into the nucleus (26). However, this is
not always the case. For example, insulin stimulates the nu-
clear translocation of several kinases, such as mitogen-acti-
vated protein kinase (27), which can then directly phosphoryl-
ate nuclear protein substrates, or activate other kinases (28).
In the present study SAPK2 is shown to be present exclusively
in the cytoplasm, suggesting that the events governing PDX1
activation occur exclusively in the cytoplasm, promoting trans-
location of an active PDX1 into the nucleus.
Glucose activation of PDX1 nuclear translocation occurs
through a
b
-cell signaling pathway involving PI 3-kinase, and
resulting in the activation of SAPK2. Hence, overexpression of
SAPK2 can artificially promote PDX1 translocation to the nu-
cleus even in lower glucose concentrations. In fact, overexpres-
sion of SAPK2 stimulated PDX1 binding activity to levels
higher than those seen in high glucose (Fig. 4). This hyper-
stimulation of PDX1 by overexpression of SAPK2 occurred in
0.5 or 20 mMglucose, and appeared to reflect activation of
PDX1 at maximum efficiency. In untransfected
b
-cells endog-
enous SAPK2 may represent a rate-limiting step in PDX1
activation, with glucose eliciting a controlled stimulation.
Overexpression of SAPK2 to artificially high levels may there-
fore result in a hyperstimulation of the PDX1 activation path-
way, resulting in an increased ability to phosphorylate and
translocate all available PDX1 protein. This “hyperstimula-
tion” of PDX1 binding activity did not appear to be reflected in
the levels of insulin gene transcription relative to those ob-
served in high glucose. This is not surprising, as PDX1 alone
does not control transcription of the insulin gene. Several other
factors (e.g. E1 and E2 site binding factors) are absolutely
required for transcription of the insulin gene. Increased PDX1
binding activity will only exert an effect in the presence of these
factors, and a maximum response would require their com-
bined activities.
Glucose stimulates PDX1 phosphorylation and nuclear
translocation through SAPK2. PDX1 is not unique in this re-
spect; CREB, which also binds to the human insulin gene
promoter (15, 16), has its binding activity controlled through
alteration in its phosphorylation status. Recent studies in SK-
N-MC cells have shown that CREB-dependent transcription is
triggered by co-transfection with SAPK2 (29), which may act in
this case through stimulation of a downstream kinase such as
MAPKAPK2 (30). However, SAPK2 has been shown to directly
phosphorylate several transcription factors, including ELK1,
CHOP, and MEF2C (17). Stress-activated protein kinases have
also been implicated in the nuclear localization of other tran-
scription factors. In EL-4 D6/76 cells, IL-1 activation of inter-
leukin-1 receptor-associated kinase and of stress-activated pro-
FIG.4.Overexpression of SAPK2 mimics the effect of high glucose on PDX1. A, Western blot analysis of untransfected MIN6 cell extract
(lanes 1–3), and extract prepared from MIN6 cells transfected with pSG-SAPK2 (lanes 4–6), using a specific anti-SAPK2 antibody. SAPK2 has an
apparent molecular mass of 38 kDa. B, Western blot analysis (anti-PDX1) of cytoplasmic (lanes 1–4) or nuclear (lanes 5–8) extracts prepared from
untransfected MIN6 cells (2), or MIN6 cells overexpressing SAPK2 (1). Samples were incubated in 0.5 mMglucose for 5 h, or 0.5 mMglucose for
5 h followed by stimulation for 30 min at 16 mMglucose, as indicated. C, electrophoretic mobility shift assay of MIN6 nuclear extracts prepared
from cells incubated in 0.5 mMglucose for 5 h (lanes 1–3), 0.5 mMglucose for 5 h followed by 30 min at 16 mMglucose (lanes 4–6), or cells
overexpressing SAPK2 which were incubated in 0.5 mMglucose for 5 h. EMSA analysis was performed using the A3 site of the human insulin gene
promoter as probe.
Nuclear Translocation of PDX1 1015
by guest on November 15, 2016http://www.jbc.org/Downloaded from
tein kinases, promotes translocation of transcription factors
NF-
k
B and IL-1 NF to the nucleus, resulting in the induction of
IL-2 mRNA synthesis (30). The phosphorylation induced nu-
clear translocation of Rel family proteins like NF-
k
B is induced
by an extraordinary large number of agents (26), cellular stress
being one example from many.
It has been proposed that glucose increases endogenous in-
sulin mRNA levels and insulin promoter activity in HIT T15
cells via exocytosis of insulin leading to autocrine stimulation
of the
b
-cell insulin receptor (31). The insulin stimulatory ef-
fects on mRNA levels were mediated through PI 3-kinase, p70
S6 kinase, and calmodulin kinase pathways. We have also
found that insulin (1–10 ng/ml) stimulates PDX1 DNA binding
activity in human islets of Langerhans and the human insulin
promoter (pGL-LUC200) activity in MIN-6 cells.
2
However, in
our study the effects of insulin were inhibited by the PI 3-ki-
nase inhibitors wortmannin (50 nM) and LY 294002 (10
m
M) and
by the SAPK2 inhibitor SB 203580 (7
m
M), but not by the p70 S6
kinase inhibitor rapamycin (50
m
M). This suggests that modu-
lation of endogenous insulin mRNA levels by insulin may in-
volve additional pathways to those described for the activation
of PDX1.
In summary, glucose stimulates translocation of the home-
odomain transcription factor PDX1 from the cytoplasm to the
nucleus in pancreatic
b
-cells via a cell signaling pathway in-
volving SAPK2. This translocation is associated with a 5-fold
increase in insulin promoter activity, and may represent a
pivotal event in the nutrient regulation of insulin mRNA
production.
Acknowledgment—We thank Prof. P. Cohen for helpful suggestions
and provision of reagents.
REFERENCES
1. Ohlsson, H., Karlsson, K., and Edlund, T. (1993) EMBO J. 12, 4251–4259
2. Miller, C. P., McGehee, R. E., Jr., and Habener, J. F. (1993) EMBO J. 13,
1145–1156
3. Leonard, J., Peers, B., Johnson, T., Ferreri, S. L., and Montminy, M. R. (1993)
Mol. Endocrinol. 7, 1275–1283
4. Boam, D. W. S., and Docherty, K. (1989) Biochem. J. 264, 233–239
5. Jonsson, J., Carlsson, L., Edlund, T., and Edlund, H. (1994) Nature 371,
606–609
6. Guz, Y., Montminy, M. R., Stein, R., Leonard, J., Gamer, L. W., Wright,
C. V. E., and Teitelman, G. (1995) Development 121, 11–18
7. Clark, A. R., Petersen, H., Read, M. L., Scott, V., Michelsen, B., and Docherty,
K. (1993) FEBS Lett. 329, 139–143
8. Petersen, H. V., Serup, P., Leonard, J., Michelsen, B. K., and Madsen, O. D.
(1994) Proc. Natl. Acad. Sci. U. S. A. 91, 10465–10469
9. Peers, B. J., Leonard, J., Sharma, S., Teitelman, G., and Montminy, M. R.
(1995) Mol. Endocrinol. 8, 1798–1806
10. Serup, P., Jensen, J., Andersen, F. G., Jorgensen, M. C., Blume, N., Holst, J. J.,
and Madsen, O. D. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 9015–9020
11. Macfarlane, W. M., Cragg, H., Docherty, H. M., Read, M. L., James, R. F. L.,
Aynsley-Green, A., and Docherty, K. (1997) FEBS Lett. 413, 304–308
12. Kajimoto, Y., Watabe, H., Matsuoka, T., Kaneto, H., Fujitani, Y., Miyazaki,
J.-I., and Yamasaki, Y. (1997) J. Clin. Invest. 100, 1840–1846
13. Melloul, D., Ben-Neriah., Y., and Cerasi, E. (1993) Proc. Natl. Acad. Sci.
U. S. A. 90, 3865–3869
14. Macfarlane, W. M., Read, M. L., Gilligan, M., Bujalska, I., and Docherty, K.
(1994) Biochem. J. 303, 625–631
15. Marshak, S., Totary, H., Cerasi, E., and Melloul, D. (1996) Proc. Natl. Acad.
Sci. U. S. A. 93, 15057–15062
16. Macfarlane, W. M., Smith, S. B., James, R. F. L., Clifton, A. D., Doza, Y. N.,
Cohen, P., and Docherty, K. (1997) J. Biol. Chem. 272, 20936–20944
17. Cohen, P. (1997) Trends Cell Biol. 7, 353–361
18. Boam, D. S., Clark, A. R., and Docherty, K., (1990) J. Biol. Chem. 265,
8285–8296
19. Ricordi, C., Lacy, P. E., and Scharp, D. W. (1989) Diabetes 38, 140–145
20. Schreiber, E., Matthias, P., Muller, M. M., and Schaffner, W. (1989) Nucleic
Acids Res. 17, 6419
21. Groom, L. A., Sneddon, A. A., Alessi, D. R., Dowd, S., and Keyse, S. M. (1996)
EMBO J. 15, 3621–3632
22. Rafiq, I., Kennedy, H. J., and Rutter, G. A. (1998) J. Biol. Chem. 273,
23241–23247
23. Duverger, E., Pellerin-Mendes, C., Mayer, R., Roche, A., and Monsigny, M
(1995) J. Clin. Sci. 108, 1325–1332
24. Mahajan, R., Delphin, C., Guan, T., Gerace, L., and Melchiar, F (1997) Cell 88,
97–107
25. Matunis, M. J., Coutavas, E., and Blobel, G (1996) J. Cell Biol. 135, 1457–1470
26. Thanos, D., and Maniatis, T (1995) Cell 80, 529–532
27. Chen, R., Sarnecki, C., and Blenis, J. (1992) Mol. Cell. Biol. 12, 915–927
28. Traverse, S., Gomez, N., Paterson, H., Marshall, C., and Cohen, P. (1992)
Biochem. J. 288, 351–355
29. Kim, J., and Kahn, C. R. (1997) Biochem. J. 323, 621–627
30. Tan, Y., Rouse, J., Zhang, A., Cariati, S., Cohen, P., and Comb, M. J. (1996)
EMBO J. 15, 4629–4642
31. Leibiger, I. B., Leibiger, B., Moele, T, and Berggren, P.-O. (1998) Mol. Cell 1,
933–938
2
H. Wu, W. M. Macfarlane, and K. Docherty, unpublished results.
FIG.5. Overexpression of SAPK2 mimics the effect of high
glucose concentrations on pGL-Luc200 reporter gene activity.
A, Western blot analysis of cytoplasmic (C) and nuclear (N) extracts
prepared from MIN6 cells overexpressing SAPK2, using a specific anti-
SAPK2 antibody. SAPK2 has an apparent molecular mass of 38 kDa. B,
MIN6 cells were transfected with the control construct pGL-Luc (LUC,
lanes 1–4), or with pGL-Luc200 (LUC200,lanes 5–8), and were incu-
bated in 0.5 mM(white)or16mMglucose (black), or 0.5 mMglucose in
the presence of 20
m
MSB 203580 (hatched). In lanes 3,4, and 7–9, cells
were co-transfected with SAPK2 as indicated.
Nuclear Translocation of PDX11016
by guest on November 15, 2016http://www.jbc.org/Downloaded from
Roger F. L. James and Kevin Docherty
Wendy M. Macfarlane, Caroline M. McKinnon, Zoe A. Felton-Edkins, Helen Cragg,
-Cellsβfrom the Cytoplasm to the Nucleus in Pancreatic
Glucose Stimulates Translocation of the Homeodomain Transcription Factor PDX1
doi: 10.1074/jbc.274.2.1011
1999, 274:1011-1016.J. Biol. Chem.
http://www.jbc.org/content/274/2/1011Access the most updated version of this article at
Alerts:
When a correction for this article is posted• When this article is cited•
to choose from all of JBC's e-mail alertsClick here
http://www.jbc.org/content/274/2/1011.full.html#ref-list-1
This article cites 31 references, 16 of which can be accessed free at
by guest on November 15, 2016http://www.jbc.org/Downloaded from
